WIT Press


Invited Paper Characterization Of Object Shapes By Singular Points: With An Application To The Feature Extraction Of Human Facial Expressions

Price

Free (open access)

Paper DOI

10.2495/VID930031

Volume

5

Pages

15

Published

1993

Size

1,147 kb

Author(s)

H. Saji, Y. Shinagawa, T.L. Kunii, H. Hioki, K. Hara, N. Asada & M. Yasumoto

Abstract

Invited Paper Characterization of object shapes by singular points: with an application to the feature extraction of human facial expressions H. Saji," Y. Shinagawa," T.L. Kunii," H. Hioki,* K. Hara,* N. Asada,* M. Yasumoto^ "Department of Information Science, Faculty of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-Ku, 113 Tokyo, Japan ^Government Industrial Research Institute, Osaka, Japan ABSTRACT The surface singular points of a function defined on an object surface rep- resents the topological structure of the object. This paper proposes and reports the implementation and testing of a new method to characterize the facial expressions by extracting the singular points of the face. As the input, the surface normals of human faces obtained by the lighting switch photometry method are used. An averaging filter is used to reduce the noise in the raw data. The experimental results demonstrate the usefulness of the proposed method to characterize faci

Keywords